Ok, we need to figure how this works together with all the auto-mirrors around 
(github/asf/maven, etc).

Are we d'accord that we must _not_ do releases on master but always create a 
temporary branch for each release (and then merge back to master after the vote 
passed) ?

LieGrue,
strub




----- Original Message -----
> From: Kristian Rosenvold <kristian.rosenv...@gmail.com>
> To: Maven Developers List <dev@maven.apache.org>
> Cc: 
> Sent: Friday, September 21, 2012 10:33 AM
> Subject: Re: Scm, Surefire, Wagon migrate to git (please check) [was Plan for 
> git migration]
> 
> You don't really "crash" anything, but if the tag is rewritten you
> have to delete the local tag to actually get the updated reference to
> the tag.
> 
> The authorative repo is correct, but there may be some clones out
> there that are unaware of the failed release and the moved tag.
> 
> I still think the benefit of moving the tag > just tagging locally
> (and hence missing tags).
> 
> Could we introduce something like subtagging ? If the official release
> tag is "xyz_plugin_1.2.3", could we have the release plugin tag the
> first version as "xyz_plugin_1.2.3_0" and just automatically advance
> the tag for each re-release ? Then we *could* have some automated step
> in the post-vote process burniung the final release number based on
> highest available subtag (and delete the subtags) ?
> 
> I suppose a similar post-vote tool could also just push the local
> release tag to the repo, so I may be complicating things.....?
> 
> Kristian
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 2012/9/20 Mark Struberg <strub...@yahoo.de>:
>>  Hi folks!
>> 
>>  The problem is the rollback.
>>  In DeltaSpike we do _not_ push now. deleting a tag on one repo is not a 
> problem, but you would crash all downstream clones as re-trying a release is 
> basically a history rewrite.
>> 
>>  In DeltaSpike I do the release locally and push the tag + release branch to 
> my github clone (or any other public repo you like).
>>  After the VOTE passes I merge it to master and push it to the ASF canonical 
> repo.
>> 
>>  Not perfect neither, but worked far better than history rewrites at least 
> ;)
>> 
>>  LieGrue,
>>  strub
>> 
>> 
>> 
>> 
>>  ----- Original Message -----
>>>  From: Kristian Rosenvold <kristian.rosenv...@gmail.com>
>>>  To: Maven Developers List <dev@maven.apache.org>
>>>  Cc:
>>>  Sent: Thursday, September 20, 2012 4:56 PM
>>>  Subject: Re: Scm, Surefire, Wagon migrate to git (please check) [was 
> Plan for git migration]
>>> 
>>>  I just checked and there is no general blocking for deleting tags. We
>>>  should definitely push as part of release plugin. Tags ar cheap.
>>> 
>>>  Kristian
>>> 
>>> 
>>>  2012/9/20 Olivier Lamy <ol...@apache.org>:
>>>>   2012/9/20 Mark Derricutt <m...@talios.com>:
>>>>>   Olivier,
>>>>> 
>>>>>   I'm not checked yet ( not thru lazyness, but thru 
> buzyness, and
>>>  shooting this email just as I think of it ), in all my local maven/git 
> projects
>>>  I set in the maven-release-plugin configuration:
>>>>> 
>>>>>   <pushChanges>false</pushChanges>
>>>>>   <localCheckout>true</localCheckout>
>>>>> 
>>>>>   to prevent any upstream/origin pushes during release ( there 
> may not BE
>>>  an origin/remote yet ). There was some discussion either on here, or in 
> JIRA (
>>>  not sure off hand which ) about making these options the default for 
> any of the
>>>  supported distributed version control systems.
>>>>> 
>>>>>   Does anyone know if this was done? Should these options be 
> mandated as
>>>  a standard practise for maven+git conversion?
>>>> 
>>>>   No release yet.
>>>>   But regarding <pushChanges>false</pushChanges>, I 
> remember we
>>>  used on
>>>>   jenkins side and everybody missed the manual step (push the tag) 
> when
>>>>   released. So we moved to true.
>>>>   BTW it's something to discuss.
>>>>   Maybe pushing the tag once release vote passed ?
>>>>   I remember some discussions on ASF infra for blocking of tag 
> deletion
>>>>   (I don't know if this has been applied on the final infra) 
> because we
>>>>   sometimes delete tags if the release vote fail.
>>>>   I can test creating a fake tag and try delete it :-)
>>>> 
>>>>> 
>>>>>   Mark
>>>>> 
>>>>> 
>>>>> 
>>>>>   On 20/09/2012, at 1:40 AM, Olivier Lamy 
> <ol...@apache.org> wrote:
>>>>> 
>>>>>>   BTW I have started a page here:
>>>>>>   http://maven.apache.org/developers/conventions/git.html
>>>>>> 
>>>>>>   we could put some git tips and more conventions.
>>>>>> 
>>>>>> 
>>>>>>   2012/9/18 Arnaud Héritier <aherit...@gmail.com>:
>>>>>>>   My remarks about these repos :
>>>>>>>   * trunk will have to be renamed master (I'm not 
> sure if
>>>  we'll go further
>>>>>>>   like using the git workflow with the develop branch to 
> keep
>>>  master as
>>>>>>>   stable as possible - In any case I would recommand to 
> follow a
>>>  convention
>>>>>>>   fix/XXXX, stable/A.B.C, feature/ZZZZ to organize 
> branches)
>>>>>>>   * A strange thing I didn't have in my company 
> repositories
>>>  I recently
>>>>>>>   migrated is the commit/tag (copy for tag) for each 
> release done
>>>  by the
>>>>>>>   release plugin. I don't know if it is a different 
> setting
>>>  in the plugin or
>>>>>>>   something cleaned by filters I applied in the SVN 
> -> GIT
>>>  migration we did
>>>>>>>   * Some strange branches to cleanup ? ASF, APACHE or by 
> user
>>>  (trygvis,
>>>>>>>   evenisse ...)
>>>>>>> 
>>>>>>>   That's all what I see for now.
>>>>>>> 
>>>>>>>   Cheers
>>>>>>> 
>>>>>>> 
>>>>>>>   On Tue, Sep 18, 2012 at 8:29 PM, Arnaud Héritier
>>>  <aherit...@gmail.com>wrote:
>>>>>>> 
>>>>>>>>   I Will test them. Will they keep these ugly URLs 
> git-wip-us
>>>  in the
>>>>>>>>   future ? I'm not a fanatic but I would prefer 
> to have
>>>  something stable
>>>>>>>>   to put in our POMs for releases. I know that we 
> don't
>>>  use often scm
>>>>>>>>   infos after the release but it is sad if we know 
> that
>>>  they'll be
>>>>>>>>   discontinued in a near future. No ?
>>>>>>>> 
>>>>>>>>   Le 18 sept. 2012 à 19:23, Olivier Lamy
>>>  <ol...@apache.org> a écrit :
>>>>>>>> 
>>>>>>>>>   Repositories has been migrated to:
>>>>>>>>>   *
>>>  https://git-wip-us.apache.org/repos/asf/maven-surefire.git
>>>>>>>>>   *
>>>  https://git-wip-us.apache.org/repos/asf/maven-wagon.git
>>>>>>>>>   * 
> https://git-wip-us.apache.org/repos/asf/maven-scm.git
>>>>>>>>>   They are currently read only.
>>>>>>>>>   Content looks good.
>>>>>>>>>   If nobody complains, I will ask to move to rw 
> mode
>>>  tomorrow. (in 24H)
>>>>>>>>> 
>>>>>>>>>   2012/9/18 Olivier Lamy 
> <ol...@apache.org>:
>>>>>>>>>>   Hi
>>>>>>>>>>   See
>>>  https://issues.apache.org/jira/browse/INFRA-5266
>>>>>>>>>>   Those tree svn paths will be readonly now 
> for
>>>  migration.
>>>>>>>>>> 
>>>>>>>>>>   2012/9/12 Olivier Lamy 
> <ol...@apache.org>:
>>>>>>>>>>>   Hi Folks,
>>>>>>>>>>>   So the vote passed, now it's time 
> for
>>>  volunteers to use their fingers
>>>>>>>>   :-).
>>>>>>>>>>>   I have started a page [1] for ETA of 
> migration
>>>  (note the Volunteer
>>>>>>>>>>>   column :-) ) and for discussion on 
> some stuff
>>>  (plugins shared)
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>>   Thanks
>>>>>>>>>>>   --
>>>>>>>>>>>   Olivier Lamy
>>>>>>>>>>>   Talend: http://coders.talend.com
>>>>>>>>>>>   http://twitter.com/olamy |
>>>  http://linkedin.com/in/olamy
>>>>>>>>>>>   [1]
>>>  https://cwiki.apache.org/confluence/display/MAVEN/Git+Migration
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>>   --
>>>>>>>>>>   Olivier Lamy
>>>>>>>>>>   Talend: http://coders.talend.com
>>>>>>>>>>   http://twitter.com/olamy |
>>>  http://linkedin.com/in/olamy
>>>>>>>>> 
>>>>>>>>> 
>>>>>>>>> 
>>>>>>>>>   --
>>>>>>>>>   Olivier Lamy
>>>>>>>>>   Talend: http://coders.talend.com
>>>>>>>>>   http://twitter.com/olamy | 
> http://linkedin.com/in/olamy
>>>>>>>>> 
>>>>>>>>> 
>>>  ---------------------------------------------------------------------
>>>>>>>>>   To unsubscribe, e-mail:
>>>  dev-unsubscr...@maven.apache.org
>>>>>>>>>   For additional commands, e-mail:
>>>  dev-h...@maven.apache.org
>>>>>>>>> 
>>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>>   --
>>>>>>>   -----
>>>>>>>   Arnaud Héritier
>>>>>>>   06-89-76-64-24
>>>>>>>   http://aheritier.net
>>>>>>>   Mail/GTalk: aherit...@gmail.com
>>>>>>>   Twitter/Skype : aheritier
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>>   --
>>>>>>   Olivier Lamy
>>>>>>   Talend: http://coders.talend.com
>>>>>>   http://twitter.com/olamy | http://linkedin.com/in/olamy
>>>>>> 
>>>>>> 
>>>  ---------------------------------------------------------------------
>>>>>>   To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
>>>>>>   For additional commands, e-mail: dev-h...@maven.apache.org
>>>>>> 
>>>>> 
>>>>> 
>>>>>   
> ---------------------------------------------------------------------
>>>>>   To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
>>>>>   For additional commands, e-mail: dev-h...@maven.apache.org
>>>>> 
>>>> 
>>>> 
>>>> 
>>>>   --
>>>>   Olivier Lamy
>>>>   Talend: http://coders.talend.com
>>>>   http://twitter.com/olamy | http://linkedin.com/in/olamy
>>>> 
>>>>   
> ---------------------------------------------------------------------
>>>>   To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
>>>>   For additional commands, e-mail: dev-h...@maven.apache.org
>>>> 
>>> 
>>>  ---------------------------------------------------------------------
>>>  To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
>>>  For additional commands, e-mail: dev-h...@maven.apache.org
>>> 
>> 
>>  ---------------------------------------------------------------------
>>  To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
>>  For additional commands, e-mail: dev-h...@maven.apache.org
>> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
> For additional commands, e-mail: dev-h...@maven.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional commands, e-mail: dev-h...@maven.apache.org

Reply via email to